AgenticSeek offers a fully local, voice-enabled AI assistant that autonomously browses the web, writes code, and plans tasks, providing a private alternative to cloud-based services. It's designed for users who want to leverage powerful AI capabilities without API costs or data privacy concerns, targeting individuals with sufficient local hardware to run large language models.

How It Works

AgenticSeek utilizes a modular agent-based architecture, allowing it to select the most appropriate "agent" for a given task, such as web browsing, coding, or file system interaction. It supports various local LLM providers like Ollama and LM Studio, as well as remote APIs. The system integrates a search engine (SearxNG) and a voice interface for hands-free operation, enabling complex task execution through a series of automated steps.

Quick Start & Requirements

• Installation: Clone the repository, set up a Python virtual environment, and run install.sh (Linux/macOS) or install.bat (Windows).
• Prerequisites: Chrome browser, ChromeDriver (version matching Chrome), Docker, Python 3.10+. For local LLM execution, a GPU with at least 12GB VRAM is recommended (14B models), with 24GB+ for 32B models.
• Setup: Configure config.ini with your chosen LLM provider (local or API) and model. Start services using ./start_services.sh or start ./start_services.cmd. Run via CLI (python3 cli.py) or web UI (python3 api.py then navigate to localhost:3000).
• Links: Website, Demo

Highlighted Details

• Fully Local Operation: All components, including LLM and speech processing, can run on user hardware.
• Autonomous Capabilities: Handles web browsing, code generation/execution, and complex task planning.
• Voice Interface: Supports speech-to-text and text-to-speech for interactive use.
• Flexible Provider Support: Integrates with Ollama, LM Studio, and various APIs (OpenAI, Deepseek, etc.).

Maintenance & Community

• Contributors: Actively seeking contributors; key maintainers include Fosowl and steveh8758.
• Community: Discord, Twitter.

Licensing & Compatibility

• License: GPL-3.0. This is a copyleft license, requiring derivative works to also be open-sourced under GPL-3.0.
• Compatibility: Commercial use or linking with closed-source projects may be restricted due to the GPL-3.0 license.

Limitations & Caveats

The project is described as "Work in Progress" and is looking for contributors. The agent routing system may not always allocate the correct agent, requiring explicit user prompts. ChromeDriver version mismatches are a common installation issue. Smaller LLM models (under 14B) are not recommended for optimal performance.